Marek Czapp , Christian Muller, Pedro Aguas Fernández ...ltces.dem.ist.utl.pt/lxlaser/lxlaser2012/upload/271...16th Int Symp on Applications of Laser Techniques to Fluid Mechanics

16th Int Symp on Applications of Laser Techniques to Fluid Mechanics Lisbon, Portugal, 09-12 July, 2012

- 1 -

High-speed Stereo and 2D PIV Measurements of Two-phase Slug Flow in a Horizontal Pipe

Marek Czapp*, Christian Muller,

Pedro Aguas Fernández, Thomas Sattelmayer

Chair of Thermodynamics, Technische Universität München, Garching, Germany

* correspondent author: [email protected]

Abstract The analysis of two-phase flow in horizontal pipes is of considerable importance in many industrial heat exchange processes like in cooling systems of high pressure water reactors. The flow regime of slug flow leads to high pressure pulsations which may damage pipes and is therefore essential to understand in the context of safety-related analysis. In the oil industry velocity profiles of the liquid phase are of immanent importance for the validation of numerical models or even new approaches in CFD simulations.

The aim of this work is to gain deeper knowledge on slug flow and especially on the slug formation process as well as the establishment of Stereo PIV regarding two-phase pipe flow investigations. In the experiments, the working fluids water and air were led into a horizontal and transparent pipe of 9.46 m length and 54 mm inner diameter. By varying the superficial fluid inlet velocities flow regimes such as stratified, wavy, slug and pipe flow were reproduced. For the first time high-speed Stereo PIV combined with LIF was successfully applied on round pipe geometry to determine the instantaneous three-component water velocity field of a slug formation process in the pipe cross-section. Furthermore, high-speed 2D PIV measurements were done with two parallel-aligned cameras to determine an axial velocity field in the axial center of the pipe.

Both PIV methods are combined with Laser Induced Fluorescence and use two synchronously operating high-speed cameras. In addition, this approach allows comparing and validating both measurement techniques in two phase flow and reveals their limitations. Since this measurement technique is non-intrusive, the slug flow itself is not affected. Spatial and temporal resolution reveal secondary flow pattern, which are about one order of magnitude smaller than the mean flow velocity. Furthermore, Stereo PIV determines flow rates, cross-sectional liquid void fraction, and axial vortices. Both techniques generate data, such as interfacial area and flow regime dependent turbulence intensities, which are important for the comparison with numerical simulations. Keywords: Stereo PIV, two-phase flow, slug flow, slug onset

1. Introduction

In the last decades, slug flow and especially the slug formation process were subjects of many

investigations, but most of the experimental approaches were phenomenological and did not

consider important fluid velocity fields. PIV in single liquid phase flow is of common practice,

whereas in Lindken and Merzkirch (2002) it was the first time that a vertical water column flow

including gas bubbles was analyzed by means of PIV combined with Laser Induced Fluorescence

(LIF). This measurement principle was applied by Carpintero-Rogero et al. (2006) to round pipe

geometry to determine two-dimensional water velocity fields in slug flow. Carneiro et al. (2011)

analyzed horizontal slug flow statistically by means of 2D PIV measurements. The results of

velocity profiles were implemented in a one-dimensional slug flow model and according

simulations were compared with the experiment. Nevertheless, appropriate literature is limited

to two-dimensional velocity fields.

Stereo PIV is able to determine flow rates of two-phase flow in pipes of industrial machines as

e.g. heat exchangers. In the oil industry the velocity profile of the fluid phase is of immanent

importance for validation of numerical one-dimensional models or even new approaches of CFD


- 2 -

simulations. Van Doornel et al. (2003) used Stereo PIV the first time to analyze the transition from

laminar to turbulent pipe flow, revealing vorticity structures. Marassi et al. (2004) measured the

pipe flow in cardiac valve prosthesis by means of Stereo PIV. Both papers show the applicability of

Stereo PIV to single-phase water pipe flow. The first time, Czapp et al. (2011) applied this

measurement technique to a round pipe geometry to determine cross-sectional three-dimensional

water velocity fields of two-phase slug flow. Furthermore, instantaneous and time averaged water

flow rates, vorticity and liquid void fraction were obtained. In a next step, Czapp et al. (2012)

successfully compared Stereo PIV results of slug flows with numerical CFD simulations. Time

normalized interface shapes and axial velocity fields, covering the pipe cross-section, were in

considerable agreement.

In this work, 2D and Stereo PIV velocity profiles will be compared regarding pipe flow test

cases. After this reference case, the slug onset and a fully developed slug will be analyzed by means

of using Stereo PIV and LIF. High spatial and temporal resolution of the three-dimensional liquid

velocity reveals vertical profiles and cross-sectional distribution of axial water. Furthermore, it

gains results of interfacial area and flow regime dependent turbulence intensities, which can be

compared with data obtained by numerical simulations.

The slug formation process is briefly described for a better comprehension. Two-phase slug

flow can appear in pipes of different diameters D and lengths L, whereas slugs fully develop in

horizontal pipes after a minimum length of 200 D. The pipe inclination, fluid viscosity and density

play an important role for the slug onset. After a certain equilibrium length small perturbations can

initiate a slug formation process. This process in a horizontal pipe is illustrated in Fig. 1. Constant

but unequal water and air inlet velocities will lead in most of the cases to wavy flow (A). Water

waves and Kelvin-Helmholtz instability lead to a water level rising and consequently to a reduction

of the cross section for air (B). The air accelerates and due to the Bernoulli effect water is sucked to

the top of the pipe (C). If the water completely fills the cross section a so called slug onset takes

place. Air compression leads to water acceleration and water is accumulated downstream, which

means the slug length increases (D). Higher fluid inlet velocities lead to gas entrainment in the front

of the plug in form of means bubbles.

Fig. 1: Slug initiation process with wavy flow (A), begin of slug onset (B), slug onset (C), and slug

expansion (D).

2. Experimental setup and Particle Image Velocimetry

All experiments on water-air two-phase flow were carried out in a two-phase flow facility with

a horizontal transparent pipe of 9.46 m length and 54 mm inner diameter as test-section (see Fig. 2).

At the beginning of the test section a two-phase mixer directs both fluids vertically separated into

the pipe. The entrance cross section is equally divided for both fluids by a thin separation plate and

the axial position z = 0 is located at its end.

Superficial fluid velocities are kept constant during every measurement, respectively volume

flow rates. The flow regime can be set by variation of these inlet velocities at the mixing section.

This allows the generation of different flow regimes, such as stratified, wavy, plug, slug, and pipe

flow. A closed water loop is used due to the use of expensive fluorescent rhodamin B coated tracer

particles. It consists of a tank, a pump and a bypass. A roots-blower, a heat exchanger, and a bypass

provide the air. Reynolds numbers of up to Re ≈ 81.000 are achieved in presented measurements

with this test facility.


- 3 -

Fig. 2: Side view sketch of the two-phase test facility with PIV measurement positions 1 and 2.

2.1 General Particle Image Velocimetry settings

The 2D and Stereo PIV setups are placed at both positions 1 and 2, as shown in Fig. 2. Slug

onset is investigated at position 1 (z = 2.22 m), whereas slug expansion is recorded at position 2 (z

= 7.26 m). In both cases a water filled box, covering the pipe section, is used to reduce the light

refraction between the camera and the fluid in the pipe (see Fig. 3). The NdYLF laser excites

fluorescent rhodamine B coated polymer tracer particles at the wavelength λlaser = 526 nm. These

particles emit light at λp,em = 573 nm that that reaches the lenses (f = 60 mm, f # = 2.8) of the high-

speed cameras through an optical high-pass filter (λcut = 570 nm).

Two high-speed cameras with a resolution of 1024 x 1024 pixels and a maximum recording

frequency of 1 kHz are used. The recording time is 1.024 seconds with the maximum spatial and

temporal resolution. Decreasing the resolution or the recording frequency increases the recording

duration.

The still warped PIV images are cross-correlated using an adaptive correlation scheme of 322

and 162 pixel interrogation areas (IA) with 50% overlap. This process is iteratively optimized using

three refinement steps and consequently gives reliable results and resolves small turbulence

structures. In order to keep the signal-to-noise ratio for 16 x 16 pixel IAs at a tolerable level, they

contain 5 - 10 particles per IA. The correlation peak detection works with high sub-pixel accuracy

of 1/64 pixel. Phantom correlations at the edge of the IA can be reduced by a weighting function

(window function), here a Gaussian window. The DC component of the FFT is removed by a

NoDC filter. A local median filter with a 3 x 3 pixel mask and an acceptance factor of 0.3 - 0.4

substitutes invalid correlations in the correlation plane.

2.2 2D Particle Image Velocimetry (2D2C PIV)

In the 2D2C PIV measurement setup both cameras are parallel aligned and directed

perpendicular to the pipe and the water filled box (see Fig. 3, left). The laser light sheet illuminates

the pipe vertically from the bottom at the axial plane x = 0. Even in case of constant superficial

velocities the slug onset is not located at a defined axial position but fluctuates with higher

Reynolds numbers. Thus, a large axial field of view is required covering the propagation of the slug

formation process. A resolution of approximately 950 x 1900 pixel of the region of interest lead to a

magnification factor of M2D = 17.6 pixel/mm, which is high enough to resolve vortices.

448 mm

448 mm


- 4 -

2.3 Stereo Particle Image Velocimetry (2D3C PIV)

Due to the characteristic shape of a developed slug it is necessary to record tracer particles in

the slug end with a camera setup upstream (A) and in the slug front downstream (B) of the light

sheet to measure (see Fig. 3, right). The reason lies in the optical access to the tracer particles in the

water phase, which are detectable when only not interrupted by the air phase. The most important

parameters and angles of this setup are listed in Tab. 1.

Tab. 1: Parameters of PIV setups.

Gen

eral

Recording method Dual frame / single exposure Optical filter High-pass: λcut = 570 nm

Recording camera APX Photron Fastcam Sampling frequency 450 - 1,000 Hz

Recording medium CMOS, 1024 x 1024 pixel Pulse Delay ∆t = 150 - 500 µs

Recording lenses f = 60 mm, f# = 2.8 Seeding material Rhodamine B polymer

Illumination Nd:YAG laser = 532 nm microspheres (dp ≈ 10.5 µm)

2D

Light sheet xy-plane, s = 1.5 mm Field of view 238 x 54 mm² (W x H)

Max. in-plane velocity w ≈ 6 m/s Interrogation area ≤ 265 x 60 (W x H)

3D

Light sheet xy-plane, s = 2 mm Dyn. spatial range DSR ≈ 34 - 68

Max. in-plane velocity vmax = 4 m/s Dyn. velocity range DVR ≈ 0.9 -1.8

Interrogation area 532 x 792 pix (W x H) Observation angle α1 ≈ α2 ≈ 41°

Field of view 54 x 54 mm² (W x H) Scheimpflug angles γ1 ≈ γ2 ≈ 12.5°

Number of vectors ≤ 63 x 90 (W x H) Scheimpf. distances a ≈ b ≈ 232 mm, c = 290 mm

Fig. 3: Experimental setup of 2D PIV (left) and Stereo PIV (right) for measurements upstream

(camera setup A) and downstream (camera setup B) the laser light sheet plane.

To calibrate a defined volume, a two-dimensional target grid has been moved in z-direction

through the water filled pipe and was recorded by both cameras at defined axial positions around

the laser light sheet. The target images should cover the position of the light section. This enables

the reconstruction of the tracer particle movement in three dimensions. Based on the target images,

Willert (2006) used a so-called pinhole model to reconstruct the world coordinates (X, Y, Z) on the

image coordinates (x, y):

��

�� ∙ ��

�� (1)

The misalignment between the calibration target and the laser light sheet leads to a disparity of

both cameras in the field of view. Figure 4 shows the disparity map before and after disparity

112 mm

Objective

Laser light sheet

2D aquarium

Cam#2Cam#1


- 5 -

correction for the calibration and Tab. 4 the shift in z direction for each correction step. A least-

squares polynomial describing the translational and rotational disparity is computed during the

disparity correction process (see Eq. 2). It returns the axial shift and rotating angles respective to X-

and Y-axis for a given disparity map. After three iterations steps the projection quality is strongly

improved. It is clearly visible, that the disparity has been reduced by the correction, especially in the

cross-sectional center. This fact can also be quantified by the mean disparity in X-direction. Only

the X-direction is considered here, because X-displacements in 2D2C displacement fields mainly

determine the axial misalignment, which is usually the error source as shown in Raffel et al. (2007).

Geometrical reconstruction of the (X,Y)-displacements give the Z-displacement. These

displacements with the displacement vector [��,��, ��,��, ��,��]� are fitted to a simple polynomial:

��,�� ! + #��,��+ $��,�� (2)

Solved for [ !, # , $]� in a least-squares sense, the translational misalignment !, the rotational

misalignment around the Y-axis #, and around the X-axis $ are obtained.

Tab. 2: Correction obtained

by Disparity map Dx

∆zkorr

[mm]

∆αkorr

[ ° ]

∆βkorr

[ ° ]

original 0.0 0. 0.0

corr. 1 1.1 -0.4 0.3

corr. 2 1.2 -0.5 0.2

corr. 3 1.2 -0.4 0.4

Fig. 4: Disparity map Dx of spatial displacement in x direction.

The measurement uncertainty results from the uncertainty of recorded particle tracers on the

camera chip. Their inaccuracy of particle displacements is indicated in the literature as 0.1 pixel.

Regarding the fixed magnification factor M, the inaccuracy of the particle displacements only

results from time displacement between two single exposures. With the reconstruction of Stereo

PIV velocity fields according to the pinhole model (see Eq. 1, C. Willert (1997)), the

inaccuracy for all three spatial directions is given as about 6.0 cm/s for horizontal and vertical, and

about 7.2 cm/s for the axial velocity component. The Contrast-Limited Adaptive Histogram

Equalization (CLAHE) filter enhances the local contrast of a PIV raw image by equalizing the

intensity histogram of a defined area around a specific pixel. A median filter treats as a high-pass

filter and removes irrelevant intensities, for example obtained by the reflection at the interface.

3. Results and comparison of 2D and Stereo PIV

The measurement of the liquid phase of a gas-liquid mixture with Stereo PIV is not a common

practice yet. Therefore, a reference case that validates the feasibility of Stereo PIV regarding two-

phase flow is recommended. The test case of pipe flow provides the measurement errors and

accuracy limits. As the current Stereo PIV setup is planned to record the liquid phase of the gas-

liquid flow, pipe flow represents the straightforward reference case.

-20 -10 0 10 20

-20

-10

0

10

20

X [mm]

Y [m

m]

-3

-2

-1

0

1

2

3

-20 -10 0 10 20

-20

-10

0

10

20

X [mm]

Y [m

m]

-3

-2

-1

0

1

2

3original correction


- 6 -

3.1 Reference cases of water pipe flows

Four pipe flows with different water superficial velocities (pipe1, pipe2, pipe3, pipe4) have

been measured by Stereo PIV at position 1 (z = 2.22 m) to determine the limits, error origins and

the quality of the results for this measurement technique applied to round pipe geometry. The

superficial inlet water velocities vary from 0.505 m/s to 1.506 m/s with an accuracy of 0,1% (see

Tab. 3, left). As an indicator for the quality of the results vertical profiles of axial velocity serve for

the comparison between three pipe flows (pipe5, pipe6, pipe7) measured with 2D PIV (see Tab. 3,

right). All measurements were recorded with a frequency of frec = 1,000 Hz and the recording

duration was adapted to the flow regime. Comparable flows of 2D and Stereo PIV are depicted in

the same line but measurements were not done at the same time and so each slug may differ from

the previous one.

Tab. 3: Parameters of test cases and flow regimes measured with 2D and Stereo PIV.

Recording time varies with test cases, whereas frequency is set to 1,000 Hz. Stereo PIV 2D PIV

Regime Case %&'� [m/s] %(�) [m/s] ∆+,-. [s] Case %&'� [m/s] %(�) [m/s] ∆+,-. [s]

Pipe flow pipe1 - 0.50 500 pipe5 - 0.51 500



Pipe flow pipe4 - 1.51 250

Slug onset slug1 1.60 0.73 250 slug3 1.60 0.66 500

Slug expan. slug2 1.04 0.71 250 slug4 1.11 0.64 500

The vertical axial velocity profiles �/01,2 of three comparable pipe flows with same inlet

velocity are depicted in figure 5 on the left side. Especially for the two slowest pipe flow cases

(%(�)≈ 0.5 and %(�)≈ 0.68) the velocity profiles and total velocities are in remarkable agreement.

Although the case with highest comparable velocity (%(�)≈ 1.06 m/s) shows slightly higher velocity

in 2D PIV, all profiles normalized to the maximum axial velocity �/2,3'0, are self-similar (see Fig.

5, middle and right).

Fig. 5: Axial velocity profiles of pipe test cases measured with 2D and Stereo PIV.

Furthermore, corresponding Reynolds numbers, measured mean and maximum axial water

velocity are depicted in the Tab. 4. It shows the temporal averaged axial velocity �/2,3'0 over the

superficial inlet velocity %(�) measured by the magnetic-inductive flow mass meter. The relative

error between both values increases with higher superficial velocities but remains within a tolerable

level of 6%. According to the Tab. 4 the relative error is not proportional to inlet velocity but tends

to increase with higher velocities.


- 7 -

Tab. 4: Pipe flow measured with Stereo PIV at axial position 1 (z = 2.22 m). Case 456 [ - ] %(�) [m/s] �/01,2 [m/s] 76 [m/s] �/2,3'0 [m/s] ∆+,-. [s]

pipe1 27,226 0.505 0.487 3.5 0.616 2.048

pipe2 36,791 0.682 0.662 3.0 0.822 2.048

pipe3 57,289 1.063 1.025 3.5 1.237 2.048

pipe4 81,194 1.506 1.413 6.2 1.717 2.048

The 2D PIV measurements reveal the axial and vertical velocity field in the vertical symmetrical

plane (x = 0, z) but are therefore not capable of measuring the volume flow rate. The axial velocity

of both measurement techniques can be compared in the intersection line of their laser light sheet at

y(x = 0, z = const.). In Fig. 6 all time-averaged profiles are shown for the first three pipe flow cases,

whereas SPIV is depicted with continuous lines and 2D PIV with dashed lines.

The time averaged mean axial velocity �/01,2 slightly varies in relation to its maximum

�/2,3'0.The maximum axial velocity of the four pipe flow cases is about 1.2 times larger than the

corresponding mean axial velocity.

�/01,2 9 ∬ �;� ;�<.=>

?<.=>

2@AB2C

;+, �/3'0 maxG�/01,2H

The highest RMS values of axial velocity are concentric distributed and reached close to the

pipe wall, where fast water from the middle of the pipe is mixed with the pipe boundary.

Fig. 6: Pipe test cases 1 to 4 from left to the right measured with Stereo PIV.

3.2 Slug formation process

The temporal development of the cross-sectional axial velocity (see Fig. 7) gives an insight into

the slug formation process of slug1 and reveals characteristic cross-sectional shape of liquid-

fraction. The slug formation process and consequently the time scale propagate from the left to the

right hand side.


- 8 -

Fig. 7: Time dependent axial water velocity w(x, y, t) and the iso-surface of vertical water velocity

(left I0.15 m/s, right I-0.15 m/s) of the formation process of slug1, measured with Stereo PIV at

a certain axial position of the pipe.

Starting with wavy flow (t ≈ 0 s), the water is rapidly filling the whole cross section and

suddenly accelerated (t ≈ 0.035 s) from 0.8 to 1.3 m/s. The red iso-surface of v(x, y, t) I 0.15 m/s

indicates the region of vertical velocity which finally leads to the water filling of the whole cross-

section. Further stream down, the front of an elongated bubble of compressed air appears at t

I 0.65, which accelerates the slug. A high negative vertical water velocity, indicated as blue iso-

surface of v(x, y, t) I-0.15 m/s from t = 0.05 - 0.1 ms, has a “horseshoe” shape and lead to the

breakdown of the slug. The velocity decreases from 1.3 to 0.9 m/s within 20 ms and ends again in

wavy flow. As expected and known from literature, the water level stream down the slug is far

lower than stream up.

Fig. 8: Slug formation process of slug1. Temporal devolution of liquid fraction and vertical velocity

(top) and contour plots of axial, vertical and horizontal velocity (middle) and axial vorticity

(bottom) for time steps A-E, G.


- 9 -

Further information of the slug formation process are given in Fig. 8 which shows the time

distributed liquid void fraction and vertical velocity of slug1 at the top and axial, vertical, horizontal

velocities and vorticity of the pipe cross section of five instantaneous time steps. Compared to the

liquid fraction (blue), the vertical velocity (red) reaches its maximum approximately 10 ms earlier,

which means the water is maximal lifted before reaching the top of the pipe. At time step D, when

the pipe is still completely filled with water, it has already negative vertical velocity.

The highest vertical velocities v(x, y, t) are also in this figure reached at the beginning (B) and the

end of the formation process (F), whereas all other time steps show no significances in v(x, y ,t).

The slug formation process of slug 3 was measured with 2D PIV the first time. Figure 9 reveals

the axial and vertical velocity at four equidistant time steps A-D (∆ti = 6 ms). At the beginning, two

water tips at the interface were generated and axial accelerated (A) by the high air velocity (here not

measured) and sucked to the top, due to the Bernoulli effect. Both of them are rising (B), but solely

the first one touches the top of the pipe (C) and is more and more developing in it´s length (D). In

all time steps the vertical water velocity is positive at the slug front and logically negative at its end.

The domain of high vertical velocity is increasing with each time step and nearly filling the whole

cross section.

Fig. 9: Slug formation process of slug3 for four equidistant time steps A-D (∆ti = 6 ms); axial

velocity w(y, z, x = 0) on the left side and vertical velocity v(y, z, x = 0) on the right side.

3.3 Expansion of fully developed slug

The expansion of slug 2 and 4 was measured with both PIV approaches at position 2 (z = 7.25

m) for same inlet conditions. Due to optical access of both cameras to the laser sheets, 2D and

Stereo PIV measurements cannot be run simultaneously, and therefore both measurements were not

done at the same time. High turbulence lead to slightly differences in each slug shape, so a variation

of each slug onset length is unavoidable. According recording parameters are shown in Tab. 3.

Figure 10 illustrates slug formation process by means of water level at x = 0 and axial velocity

profiles, independently obtained by both measurement techniques. Both time scales are

independently normalized by the rise and slope of the water but are of nearly same length.

2D2C and 2D3C results show same results for all time steps. First time step A starts with simple

stratified flow with water levels around 40%. At step B the slug formation starts with the

acceleration of water and a “splash” at the top of the pipe, seen by both approaches. This spritzer is

known from former experimental measurements and the literature and is a realistic characteristic of


- 10 -

a slug front. Time steps C - F illustrate pipe flow velocity profile for about 20 ms, whereas the

maximum axial velocity sinks from the top (C) over the middle (D) to the bottom (F). Both water

levels end at time step G again in stratified flow.

Fig. 10: Temporal devolution of expanding slugs 4 (2D2C) and 2 (2D3C) from left to the right A-G;

liquid height h(x = 0, y = 0) at the top, axial velocity profile in the middle and 2D3C contour plot of

axial velocity in the cross section.

4. Summary and conclusion

With this work, Stereo PIV starts to become established in optical whole-field measurement

techniques for two-phase slug flows resolving three-dimensional water velocities in pipe cross

sections. It has been shown that 2D and also Stereo PIV are considerable measurement techniques

in two-phase flow, whenever optical access from cameras to the liquid of interest is ensured.

Furthermore, the applicability of Stereo PIV on round pipe geometry including measurements of

two-phase slug formation is shown. Pipe flow test cases have been run as a reference case for the

comparison between 2D and Stereo PIV gained axial velocity fields. Their profiles are in

remarkable agreement for all test cases. In addition, the error of the water flow rate between

measurements with the pump and SPIV is less than 6%.

Both PIV approaches show insights into the slug formation process by means of interface

detection and axial velocity fields. 2D and Stereo PIV results are comparable and in considerable

agreement. Stereo PIV determines three-dimensional water velocity, liquid fraction and vorticity,

whereas 2D PIV complements axial resolution and resolves coherent flow structures.

The authors are grateful for the

financial support of the German

Federal Ministry of Economics

and Technology (BMWi) that

founded the research project under

the project number 1501359.

Responsibility for the content of

this paper lies with the author.


- 11 -

Carneiro J N E, Fonseca Jr R, Ortega A J, Chucuya R C, 2011. Statistical Characterization of Two-

Phase Slug Flow in a Horizontal Pipe. , XXXIII, pp.251-258.

Carpintero-Rogero E, Kröss B, Sattelmayer T, 2006. Simultaneous HS-PIV and shadowgraph

measurements of gas-liquid flows in a horizontal pipe. 13th Int Symp on Applications of Laser

Techniques to Fluid Mechanics, (Lisbon), pp.26-29.

Czapp M, Utschick M, Rutzmoser J, Sattelmayer T, 2012. Investigations on Slug Flow in a

Horizontal Pipe using Stereoscopic Particle Image Velocimetry and CFD Simulation with

Volume of Fluid. Proceedings of the 2012 20th International Conference on Nuclear

Engineering. Anaheim, USA.

Czapp M, Utschick M, Sattelmayer T, 2011. Stereo PIV Investigations on Slug Flows in a

Horizontal Pipe. 19. Fachtagung Lasermethoden in der Stroemungsmesstechnik., GALA

van Doornel C W H, Hof B, Lindken R, Westerweel J, 2003. Time Resolved Stereoscopic PIV in

Pipe Flow. Visualizing 3D Flow Structures. 5th International Symposium on Particle Image

Velocimetry, Paper 3132, pp.1-11.

Lindken R, Merzkirch W, 2002. A novel PIV technique for measurements in multiphase flows and

its application to two-phase bubbly flows. Experiments in Fluids, 33, pp.814-825.

Marassi M, Castellini P, Pinotti M, Scalise L, 2004. Cardiac valve Prosthesis Flow Performances

measured by 2D and 3D-stereo particle image Velocimetry. Experiments in Fluids, 36, pp.176-

186.

Raffel M, Willert C, Wereley S, Kompenhans J, 2007. Particle Image Velocimetry, Springer-Verlag

Berlin Heidelberg New York.

Willert C, 1997. Stereoscopic digital particle image velocimetry for application in wind tunnel

flows. Meas. Sci. Technol. 8, 1465.

Willert C, 2006. Assessment of Camera Models for Use in Planar Velocimetry Calibration.

Experiments in Fluids, 41, pp.135-143.

Documents

Marek Czapp , Christian Muller, Pedro Aguas Fernández ...ltces.dem.ist.utl.pt/lxlaser/lxlaser2012/upload/271...16th Int Symp on Applications of Laser Techniques to Fluid Mechanics